Complex Cut Out from a Photo...

I had an idea and opened the image in Select > Select and Mask.

In there, the first step was to lay down a baseline selection with the Quick Selection tool, mostly to mark non-ambiguous areas such as the tree trunk and branches.

Next, I switched to the Refine Edge Brush tool and painted the ambiguous edges like the fine branches and tiny leaves. After a minute or so, this worked a little better than I thought it would. This tool is mostly designed for mask edges where there is hair or fur, I don’t know how much Adobe tested it with trees and leaves.

But my result is still a big mess if you look closely, and I don’t know how much time it would take to create a result you would actually be happy with, or if that can be achieved.

And if you aren’t experienced with Select and Mask, it would take additional time to understand what exactly each tool is for. With Select and Mask it’s very important to use the right tool for the right purpose, because painting the wrong thing with the wrong tool can quickly ruin large areas of the mask. The main principles are, use most of the Select and Mask tools to mark un-ambiguous areas that should not be masked and are definitely not edges, and then use the Refine Edge Brush tool specifically to mark ambiguous edges…fine details, or areas that should end up along semi-transparent edges in the final mask. Never drag the Refine Edge Brush tool over an area that should be fully within the solid white or solid black areas of the mask, it is only for ambiguous edges.

Successful use of Select and Mask also requires understanding its different View modes. For the Quick Selection tool I could more easily see what I was doing with the red overlay View mode, and when checking the result I found it easier to use the black matte View mode.

Ooooh, that is a tough example. No, I don’t think the tools in Lightroom Classic would help, not even the AI-powered selections are going to properly separate the fine branches from the very similar foliage in the background.

The only thing I can think of that might speed up selection is if the original image was taken with a phone or other camera that records a depth map with the picture. If the depth map is sufficiently detailed, then in theory, the tree could be separated from the background by making a selection based on distance from camera.

But even if the camera did record a depth map, I have a feeling it wouldn’t be detailed enough to separate out the very fine branch ends.

This is going to be almost impossible without manually using the pen and making a complex selection, especially on the right side.